CN107290653B - Detection device and method for identifying PCB based on broadband magnetic induction channel characteristics - Google Patents
Detection device and method for identifying PCB based on broadband magnetic induction channel characteristics Download PDFInfo
- Publication number
- CN107290653B CN107290653B CN201710336742.5A CN201710336742A CN107290653B CN 107290653 B CN107290653 B CN 107290653B CN 201710336742 A CN201710336742 A CN 201710336742A CN 107290653 B CN107290653 B CN 107290653B
- Authority
- CN
- China
- Prior art keywords
- detection
- pcb
- module
- magnetic
- broadband
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/28—Testing of electronic circuits, e.g. by signal tracer
- G01R31/302—Contactless testing
- G01R31/308—Contactless testing using non-ionising electromagnetic radiation, e.g. optical radiation
- G01R31/309—Contactless testing using non-ionising electromagnetic radiation, e.g. optical radiation of printed or hybrid circuits or circuit substrates
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Electromagnetism (AREA)
- Toxicology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geophysics And Detection Of Objects (AREA)
- Investigating Or Analyzing Materials By The Use Of Magnetic Means (AREA)
Abstract
The invention discloses a detection device and a detection method for identifying a PCB (printed Circuit Board) based on broadband magnetic induction channel characteristics. The device comprises a detection front end responsible for receiving and transmitting broadband detection signals and a data processing rear end which performs post-processing on received signals to obtain detection results, wherein the detection front end and the data processing rear end are connected in a wired/wireless mode. The device also has an infrared counting function, and can realize a batch detection function by means of a mechanical automation module. The method obtains a spatial distribution diagram (magnetic diagram for short) of magnetic induction intensity characteristics of a detected PCB through channel inversion by means of receiving and transmitting broadband electromagnetic waves, realizes the defect detection of the PCB through establishment of a magnetic diagram information feature library and feature comparison, and realizes the function of batch detection by adopting a mechanical automation module. The broadband electromagnetic detection signal adopted by the invention has rich information, and has the advantages of high detection precision, high speed, simple operation, no harm to human body and the like.
Description
Technical Field
The invention relates to the technical fields of electromagnetics, PCB detection and the like, in particular to a detection device and method for identifying a PCB based on broadband magnetic induction channel characteristics.
Background
With the improvement of the integration degree of a printed circuit board (Printed Circuit Board, abbreviated as PCB), the popularization and application of the small-sized packaging technology, and the PCB detection plays an important role in ensuring the product quality. Currently, the PCB inspection technologies commonly used mainly include Flying Probe Testing (FPT), optical inspection (AOI), electrical testing, and X-ray inspection (AXI).
The flying probe test and the optical detection can only be used for surface detection, and can only detect the problems of burrs, circuit breaks and the like on the surface of the PCB bare board, and the problems of the defects of internal short circuit, circuit break, burrs and the like of the PCB cannot be judged. The flying probe test belongs to a contact test, and has the problems of inaccurate probe positioning and secondary damage; the detection precision of the optical detection has high requirements on the positioning accuracy of the camera. The electrical test has long feedback time, large investment in the early stage and difficult upgrading, and also belongs to contact test, and the detection process needs to contact the PCB for a long time, thus easily causing the damage of the PCB. Compared with the three methods, the X-ray detection has larger detection range and higher precision, but the X-rays have radiation pollution and can cause damage to human bodies.
Disclosure of Invention
The invention mainly aims to solve the problems that the existing PCB defect detection has long detection period and low detection precision, and defects in the PCB such as burrs, short circuits, open circuits and the like cannot be detected, and provides a detection device and a detection method for identifying the PCB based on broadband magnetic induction channel characteristics by utilizing the difference of magnetic permeability of different materials. The invention adopts the array transmitting coil, takes the broadband electromagnetic wave signal as the transmitting signal, takes the PCB to be detected as the transmission channel of the electromagnetic wave, and realizes the space identification of the magnetic channel transfer function of the detection point through the analysis and the channel inversion of the array receiving signal. By adopting the array to send and receive signals, the geometric position of the transmitting coil and the receiving signals can be utilized to accurately position and detect the magnetic channel characteristics of a certain detection point of the PCB, and then the spatial distribution of the magnetic channel transfer function of the PCB in the whole detection area is obtained through the fusion of the multi-point magnetic channel characteristics, namely, a body transfer function describing the magnetic characteristics of different positions of the PCB, and the defect detection is realized through the characteristic analysis and the comparison. The non-contact detection device and the non-contact detection method can detect the defects in the PCB, and the adopted broadband electromagnetic detection signal is rich in carrying information, and has the advantages of high detection precision, high speed, simplicity in implementation, no harm to human bodies and the like.
The invention is realized at least by one of the following technical schemes.
A detection method based on broadband magnetic induction channel feature identification PCB takes broadband electromagnetic waves as emission signals, the detected PCB is taken as a channel, transfer functions of magnetic channels where the PCB is located at detection points are inverted by utilizing array receiving signals and array emission signals, magnetic induction intensities of all points of the detected PCB are obtained, and then a spatial distribution diagram, namely a magnetic map, of the magnetic induction intensity characteristics of the detected PCB is obtained through splicing and fusion.
Further, in the detection method, during the first detection, a data packet containing the simulated feature magnetic pattern, the related feature parameters and the classification labels of the PCB is imported into a database, and a feature library is established.
Further, in the detection method, for suspicious results, a manual detection method is adopted to identify the PCB defects, and the automatic detection capability is improved through magnetic pattern feature extraction and sample training.
The invention provides a detection device for identifying a PCB based on broadband magnetic induction channel characteristics, which comprises a detection front end and a data processing rear end. The detection front end is responsible for transmitting and receiving broadband electromagnetic waves, the detected PCBs are counted by adopting an infrared technology, and the data processing rear end is responsible for carrying out post-processing on the received signals to obtain detection results. The two parts transmit data and control information through a communication module, and a wired/wireless mode can be adopted.
Further, the detection front end of the detection device comprises a broadband signal generation module, a preprocessing module, a transmitting coil array, a receiving buffer module, an infrared induction counter, a mechanical automation module, a communication module and a slave controller. The broadband signal generation module is mainly used for generating broadband detection signals, and the center frequency of the broadband detection signals can be comprehensively considered and selected according to the detection depth and resolution requirements; the preprocessing module comprises a digital-to-analog conversion unit and a power amplification unit and is responsible for converting a digital signal into an electric signal; the transmitting coil array consists of a plurality of transmitting coils and transmits broadband electromagnetic wave signals under the action of electric signals; the receiving coil array is composed of a plurality of coils like the transmitting coil array and is used for receiving electromagnetic waves passing through the detected PCB channels at different positions; the receiving buffer module comprises an analog-to-digital conversion unit and a buffer unit, and is responsible for converting a received analog signal into a digital signal and storing the digital signal; the infrared induction counter is mainly used for counting the detected PCBs and is convenient for storage; the mechanical automation module comprises a belt transmission device, a mechanical arm and other industrial automation devices and is mainly used for automatically placing the PCB so as to reduce the investment of labor cost; the communication module is mainly used for realizing communication with the data processing rear end; the slave controller is mainly used for carrying out parameter setting on each module in the detection front end and carrying out coordination control on the work among each module.
Further, the data processing rear end of the detection device comprises a man-machine interaction module, a database, a channel inversion and fusion module, a characteristic comparison module, a data storage module, a communication module and a main controller. The human-computer interaction module is mainly used for configuring parameters of the whole device system, importing the feature data detected for the first time and carrying out information interaction of manual detection; the database is used for storing a characteristic library and comprises a characteristic magnetic diagram of the PCB to be detected, related characteristic parameters and characteristic labels; the channel inversion and fusion module inverts the channel of the PCB at the detected point by utilizing the transmitting signal and the receiving signal to obtain the magnetic induction intensity of the PCB at the detected point, and then the magnetic induction intensity characteristic distribution map of the detected PCB is obtained by splicing and fusing the magnetic induction intensities of all points; the feature comparison module is used for extracting the features of the detected PCB magnetic patterns, comparing the features with the features in the PCB feature library of the same type in the database to obtain a classification result of the features, and judging the cause and the position of the defect; the data storage module is used for storing the detection result and the corresponding detection count number of the detected PCB; the communication module is mainly used for realizing communication with the detection front end; the main controller is mainly used for setting parameters of the system, and is used for coordinately controlling the function realization of each module and the mutual calling among the modules.
A detection method for identifying a PCB based on broadband magnetic induction channel characteristics comprises the following steps:
and 1, setting parameters of the device. And setting parameters through a man-machine interaction module at the rear end of the data processing. The control parameters are communicated to the slave controller via the communication module. The set parameters comprise the center frequency and the corresponding bandwidth of the broadband signal, related parameters of signal processing in the preprocessing module, state parameters of a transmitting coil array and a receiving coil array, belt transmission parameters in the mechanical automation module, movement parameters of the mechanical arm and parameters of the communication module.
And 2, automatically detecting the state of the device. Checking whether the communication connection among all modules of the automatic detection device is good, and whether the working state of all modules is normal.
And step 3, importing a simulation data packet to establish a feature library. It is determined whether the model of PCB is first detected. If not, directly executing the step 4; if yes, execute this step. And importing the characteristic magnetic patterns, the characteristic parameters of the related magnetic patterns and the characteristic labels of the PCB of the model, which are generated by the simulation system, into a database, and initially establishing a characteristic library of the PCB of the model.
And 4, emitting broadband detection electromagnetic waves. The broadband signals obtained according to the parameter setting in the step 1 are converted into electric signals through the signal processing of the preprocessing module to control the transmitting coil array, the required broadband detection electromagnetic waves are generated, and the control parameters of the transmitting coil array are set in the step 1.
And 5, receiving an electromagnetic wave signal. The electromagnetic wave passing through the channel where the detected PCB is located is received by a receiving coil array, the parameters of the receiving coil array are set by the step 1, the received information is converted into digital information through an analog-to-digital conversion unit in a receiving buffer module and is stored in a buffer unit of the module, and the received electromagnetic wave information can be transmitted to the data processing back end through a communication module.
And 6, inverting and fusing the PCB magnetic channel characteristics. And the channel inversion and fusion module inverts the magnetic channel characteristics of the PCB at the detection point to obtain the magnetic induction intensity characteristics of the PCB at the detection point, and finally obtains the magnetic diagram of the PCB by splicing and fusing the magnetic induction intensity distribution characteristics at different positions.
Step 7, magnetic channel characteristic comparison and classification: the feature comparison module directly calls the magnetic patterns in the PCB feature library of the same type in the database and corresponding feature parameters to compare, classify and judge, and the results can be divided into two types:
A. and determining a PCB detection result. The device can classify and judge the PCB through the characteristic comparison of the magnetic patterns, and the defect type and the position of the PCB are obtained. In this case, the detection result is directly recorded.
B. Suspicious PCB detection results. The device can not accurately classify and judge the characteristics of the PCB magnetic diagram through the characteristic comparison of the magnetic diagram. The PCB is marked in this case and a warning is given. When the device is idle, manual detection or repeated detection can be selected. During manual detection, not only is the detection result of the PCB required to be judged, but also feature extraction is carried out on the magnetic diagram of the detection PCB, feature labels are established and put into a database, sample features in the database are enriched, then a statistical pattern recognition module, a neural network, a support vector machine and other machine learning algorithms are adopted, the classification judgment capacity is improved through training, and the judgment accuracy in an automatic detection mode is improved.
And 8, detecting the next PCB. Judging whether an instruction for ending the detection exists, and if so, ending the automatic detection. If not, the infrared counter counts up by 1, then the next PCB detection is carried out, and the steps 4-7 are executed.
The detection device and the method for identifying the PCB based on the broadband magnetic induction channel characteristics have the following technical advantages:
(1) The broadband signal detection is adopted, so that the frequency response characteristic of the detection point can be obtained, the characteristics are rich, and the detection precision is high;
(2) The internal defects can be accurately positioned and judged by adopting array emission and array reception, and places which cannot be seen by naked eyes can be detected;
(3) The non-contact detection method is adopted, and no electrified detection is needed, so that the detection PCB is not damaged;
(4) The testing device is simple, has strong mobility, and does not need a ray shielding chamber during X-ray detection.
Drawings
FIG. 1 is a schematic diagram of a front-end magnetic tunnel for detection according to the present invention;
FIG. 2 is a schematic plan view of a transmit coil array and a receive coil array of the present invention;
FIG. 3 is a block diagram of the apparatus of the present invention;
FIG. 4 is a flow chart of the device detection of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but embodiments of the present invention are not limited thereto.
According to the requirements of PCB quality detection and combining an information theory principle and an electromagnetic theory, the invention provides a detection device and a detection method for identifying the PCB based on broadband magnetic induction channel characteristics so as to meet the requirements of modern PCB quality detection.
As shown in fig. 3, the detection device for identifying the PCB based on the characteristics of the wideband magnetic induction channel includes a detection front end and a data processing rear end, wherein the detection front end is responsible for transmitting and receiving the wideband electromagnetic wave, and the data processing rear end is responsible for post-processing the received signal to obtain a detection result; the detection front end comprises a slave controller, and a broadband signal generating module, a preprocessing module, a transmitting coil array, a receiving buffer module, an infrared induction counter, a mechanical automation module and a communication module which are respectively connected with the slave controller, wherein the broadband signal generating module, the preprocessing module and the transmitting coil array are sequentially connected, and the receiving coil array is connected with the receiving buffer module; the broadband signal generation module is mainly used for generating broadband detection signals, and the center frequency of the broadband detection signals is comprehensively considered and selected according to the detection depth and resolution requirements; the preprocessing module comprises a digital-to-analog conversion unit and a power amplification unit and is responsible for converting a digital signal into an electric signal; the transmitting coil array consists of a plurality of transmitting coils and transmits broadband electromagnetic wave signals under the action of electric signals; the receiving coil array consists of a plurality of coils and is used for receiving electromagnetic waves passing through the detected PCB channels at different positions; the receiving buffer module comprises an analog-to-digital conversion unit and a buffer unit, and is responsible for converting a received analog signal into a digital signal and storing the digital signal; the infrared induction counter is mainly used for counting the detected PCBs and is convenient for storage; the mechanical automation module comprises a belt transmission device and a mechanical arm and is mainly used for automatically placing and conveying the PCB; the communication module is mainly used for realizing communication with the data processing rear end; the slave controller is mainly used for carrying out parameter setting on each module in the detection front end and carrying out coordination control on the work among each module;
the data processing rear end comprises a database, a main controller, a man-machine interaction module, a channel inversion and fusion module, a characteristic comparison module, a data storage module and a communication module, wherein the man-machine interaction module, the channel inversion and fusion module, the characteristic comparison module, the data storage module and the communication module are respectively connected with the main controller; the human-computer interaction module is mainly used for configuring parameters of the whole device, importing feature data detected for the first time and carrying out information interaction of manual detection; the database is used for storing a characteristic library and comprises a characteristic magnetic diagram of the PCB to be detected, related characteristic parameters and characteristic labels; the channel inversion and fusion module inverts the channel of the PCB at the detected point by utilizing the transmitting signal and the receiving signal to obtain the magnetic induction intensity of the PCB at the detected point, and then the magnetic induction intensity characteristic distribution map of the detected PCB is obtained by splicing and fusing the magnetic induction intensities of all points; the feature comparison module is used for extracting features of the detected PCB magnetic patterns, comparing the feature extraction with features in a PCB feature library of the same type in the database to obtain a feature classification result, and judging the cause and the position of the defect; the data storage module is used for storing the detection result and the corresponding detection count number of the detected PCB; the communication module is mainly used for realizing communication with the detection front end; the main controller is mainly used for setting parameters of a system, realizing the coordination control of functions of each module and calling the device among the modules mutually to carry out inversion splicing and fusion on the magnetic channels at each detection point so as to obtain a magnetic induction intensity characteristic distribution diagram, namely a magnetic diagram, of the detected PCB. In a PCB, if a short circuit occurs at a certain position of a certain layer, it will be different from the dielectric distribution of a qualified PCB, and the dielectric distribution will also be different due to the short circuit at different positions. It is because of the difference of medium distribution that the detection of this device can obtain different magnetic channel characteristics, can judge whether PCB is qualified from this, and according to classification and the judgement result of characteristic, obtain the reason and the position of defect.
In this embodiment, the mechanical automation module in the device of the invention uses a belt as the placement surface for the PCB 4 to be inspected, by means of existing industrial automation. An infrared sensing counter 5 is arranged at two sides of the front belt of the magnetic field detection area, and the positions are marked in fig. 1. The mechanical automation module comprises a belt transmission device and a mechanical arm, and is mainly used for automatic placement and transmission of PCBs. The belt 6 of the belt drive rotates to carry the PCB into the detection area and the detected PCB is counted by infrared sensing. And the detection result and the count number of the detected PCB are put into a data storage module, so that the cause and the position of each defective PCB can be conveniently searched after the detection is finished. When the last PCB is detected, the mechanical automation module controls the mechanical arm to place the next PCB to be detected on the belt, controls the belt roller shaft to rotate, places the PCB to be detected in the detection area, and simultaneously the counter automatically counts.
The device structure of the present invention is shown in fig. 3. The device mainly comprises a detection front end and a data processing rear end. The communication modules can adopt the existing wired or wireless mode to realize the transmission of data and control information.
The main function of the detection front end in this particular embodiment is to take charge of the transmission and reception of broadband electromagnetic waves and to count the PCBs to be detected. The frequency of the broadband electromagnetic wave signal at the detection front end can be controlled in a low-frequency area which is easy to obtain, because the long wave has the advantages of reduced transmission attenuation, weak interference and stable signal. The frequency range is selected to be 20KHz-25KHz in this embodiment. The main controller indirectly sets parameters of the broadband signal generation module through the slave controller to enable the parameters to generate proper broadband digital signals, and the signals are converted into broadband orthogonal electric signals through a digital-to-analog conversion unit and a power amplification unit in the preprocessing module. The transmitting coil array transmits broadband electromagnetic waves under the action of the electric signal. The electromagnetic wave passes through the PCB to be detected and is then received by the receiving coil array, the electromagnetic wave information received by the receiving coil from each detection point is converted into digital information through an analog-to-digital conversion unit in the receiving buffer module, the information is stored in the buffer unit in the module, and the stored information is sent to the data analysis processing back end for analysis through the communication module. After one PCB is detected, the mechanical automation module is called, the mechanical arm is controlled to place the next detection PCB on the belt, the belt roller shaft is controlled to rotate, the detected PCB is transmitted to the detection area, and the infrared induction counter automatically counts up by 1 at the moment, so that the next PCB is detected. The parameters of the broadband detection signal, the parameters of the transmitting coil array and the receiving coil array and the parameters of the mechanical automation module are directly set and adjusted by the slave controller, and the mutual calling among the modules at the front end of the detection is also adjusted and controlled by the slave controller. The communication module is responsible for communicating with the data processing back-end data and control information. The power supply is responsible for supplying power to each module of the detection front end of the device.
The data processing rear end in the specific embodiment is mainly responsible for post-processing the received signal to obtain a detection result. The data post-processing process is that a received signal is transmitted to a channel inversion and fusion module through a communication module, the module inverts a magnetic channel at a detection point by utilizing a transmitting signal and the received signal to obtain the magnetic induction intensity of a PCB at the detection point, and then a distribution diagram of the magnetic induction intensity in the whole detection area, namely a magnetic diagram, can be obtained by splicing and fusing the magnetic induction intensity of each detection point. And the magnetic pattern is transmitted to a characteristic comparison module which is responsible for carrying out characteristic extraction and characteristic comparison on the magnetic pattern, and classification judgment is carried out according to the corresponding comparison result to obtain the detection result of the PCB to be detected. The classification judgment of the feature comparison module can adopt a statistical pattern recognition module, a neural network or a machine learning algorithm related to a support vector machine and the like to learn the feature samples, so that the classification judgment capability of the device is improved. The device stores the judging result of the detected PCB and the counting number thereof into the data storage module.
In this embodiment, the modes of PCB quality detection fall into two categories: one is an automatic detection mode and one is a manual detection mode. In the automatic detection mode, the device compares and classifies the characteristics of the sample in the characteristic library with the obtained magnetic pattern characteristics of the PCB to be detected, and automatically obtains a detection result. By means of the mechanical automation module, the device can realize batch automatic detection of PCBs under the condition of no manual monitoring. When the device can not accurately conduct classification judgment on the PCB detection of a certain model in an automatic detection mode and a suspicious judgment result appears, manual detection is adopted. At this time, the manual detection is not only needed to detect and judge the model PCB, but also needed to extract the characteristics of the magnetic pattern of the detected PCB and establish a characteristic label to be put into a database to serve as a characteristic sample, namely, the label is artificially established for the newly detected PCB magnetic pattern, such as the non-ideal states of short circuit, circuit breaking, virtual welding and the like, the characteristics of the sample are extracted, and the characteristic library of the magnetic pattern is perfected. When the number of the samples in the feature library is sufficient, the device can improve the capability of classification judgment and the detection accuracy through sample training.
The feasibility and the specific principle of the invention are as follows:
it is known from the ampere theorem that a changing electric field generates a magnetic field, and the magnetic flux generated by the magnetic field in a unit curved surface is the magnetic field strength H. After the applied magnetic field H penetrates the medium, the medium is influenced by H to generate some additional field such that the magnetic field at this point is no longer H. This process of being influenced by an external magnetic field such that the medium also has a magnetic field inside is called "magnetization". And the total magnetic field at this point, i.e. the magnetic induction B, is the vector sum of the applied magnetic field H and the additional field generated after the medium has been magnetized. Since different media have different influences on the magnetization, the magnetic induction B after passing through different media surfaces is different. The relationship between the magnetic induction intensity B and the magnetic field intensity H can be described by magnetic permeability, and the relationship expression is: b=μh, where μ is the permeability of the medium, and is related to the type, position, etc. of the medium. In the invention, the broadband signal is converted into the electric signal after pretreatment, and the electric signal acts on the transmitting coil array to obtain the broadband electromagnetic wave, so that the required broadband magnetic signal can be obtained by adjusting the parameters of the broadband signal. In the multi-layer PCB, different positions, different circuit distributions and different circuit board layers can make the medium distribution or the medium type of the PCB different, so that the broadband magnetic signal can generate different magnetic induction intensities B at different positions after passing through the PCB to be detected.
As is known from faraday's electromagnetic induction theorem, a varying magnetic field produces an electric field. The magnetic flux passing through the PCB and the magnetic flux of the receiving coil are different due to the magnetic permeability difference of materials at different positions of the PCB, so that electromotive forces generated on the receiving coil are different. By analyzing the electromotive force, the signal intensity of the received electromagnetic wave at the detection point can be obtained. And inverting the magnetic channel at the detection point by using the transmitting signal and the receiving signal to obtain the magnetic induction intensity of the PCB at the detection point, and then splicing and fusing the magnetic induction intensity of each detection point to obtain a characteristic distribution diagram of the magnetic induction intensity in the whole detection area, namely a magnetic diagram. The transmitting signal adopted by the invention is a broadband magnetic signal, the magnetic channel characteristic of the PCB at the detecting point under different frequencies, namely the frequency response characteristic, can be converted into the time domain characteristic through Laplace or Fourier inverse transformation, and the time domain characteristic of the channel is obtained. In a multilayer PCB, if a short circuit occurs at a certain position of a certain layer, the dielectric distribution at that position must be different from that of a qualified PCB, and the difference in dielectric distribution will cause inconsistent magnetic induction intensity at that position, and the distribution of magnetic induction intensity characteristics generated in the case of a short circuit at different positions will also be different. By adopting array signal transmission and array signal reception, a certain detection point in the PCB can be accurately positioned and detected by utilizing the geometric position of the transmitting coil and the receiving signal, and more accurate magnetic induction intensity and spatial distribution thereof are obtained. If a feature library of the PCB is established, the library comprises a magnetic induction intensity characteristic distribution diagram, corresponding feature parameters and feature labels of the defective PCB, and the defects, the defect reasons and the positions of the defects can be detected through feature comparison. Similarly, if the PCB is unqualified, such as circuit breaking, burrs and the like, the obtained magnetic induction characteristic distribution of the PCB is inconsistent with that of the qualified PCB, and the PCB presents corresponding defect characteristics. Therefore, the invention can detect the cause and the position of the PCB defect.
When the device executes PCB detection, whether the detection of the type PCB is performed for the first time is judged. If the PCB is detected for the first time, a data packet containing the characteristic magnetic pattern simulated by the PCB to be detected, related characteristic parameters and characteristic labels is required to be imported, and a characteristic library of the PCB with the model is established. The characteristic magnetic diagram is generated by adopting a special simulation platform, and the magnetic diagram simulation process on the simulation platform comprises the following steps: firstly, modifying a qualified circuit diagram of the PCB with the model according to the position of a common defect to obtain a defective PCB circuit diagram; then adopting the same broadband electromagnetic wave and detection method as the detection front end to simulate the detection process of the PCB qualified circuit and the defect circuit, namely obtaining a distribution diagram of the magnetic induction intensity characteristic of the whole circuit through inversion and fusion of magnetic channels of detection points of the circuit, and realizing magnetic diagram simulation; finally, extracting features from the simulated magnetic patterns and establishing feature labels, and putting the simulated magnetic patterns, the extracted features and the related feature labels into a data packet so as to conveniently import feature data. When the simulation is performed, the existing magnetic distribution of the detection environment can be simulated, so that the detection characteristics reflected by the magnetic diagram generated by the simulation system have high matching degree with the actual detection characteristics, and the detection accuracy can be improved. After the simulation data packet is imported, the PCB with the model can be automatically detected in small batches, and the automatic detection condition of the device is checked. If the automatic detection of the PCB with the model can not judge the result of the classification of the detected PCB, the manual detection is needed. When in manual detection, not only the type and the kind of the defects of the detected PCB need to be judged, but also the characteristics of the magnetic diagram of the detected PCB need to be extracted, and the labels need to be established for improving the characteristic library. After manual detection, the device can adopt algorithms such as a statistical pattern recognition module, a neural network or a support vector machine to carry out classification learning on the feature samples obtained by the detection, and the capability of the device for automatically comparing and classifying and judging the magnetic pattern features of the detected PCB is improved. The invention also adopts an infrared induction counting method to count each detected PCB. And the detection result of the detected PCB and the corresponding count number are put into the data storage module by utilizing the count number of the PCB, so that the unqualified reasons of the detection of different PCBs can be conveniently searched.
A schematic diagram of a detection magnetic tunnel of the present invention is shown in fig. 1. The magnetic induction lines 3 of the detection region are indicated by broken lines in fig. 1. In this embodiment, the transmitting coil array 1 and the receiving coil array 2 are each composed of a plurality of coils, wherein the receiving coils may be replaced by magnetic sensitive elements such as hall elements, magnetic sensitive resistors, magnetic sensitive diodes, and the like, and their layout is shown in fig. 2. The receiving coils are arranged near the belt to receive electromagnetic wave signals conveniently, and the centers of the coils 201 are respectively provided with a magnetic core 202 (such as iron and the like) for gathering magnetic flux to enable magnetic induction lines in the middle position to be dense so as to strengthen the transmission and the reception of the magnetic induction signals. By selecting proper broadband detection signals, the signals are preprocessed into electric signals, and the electric signals act on the transmitting coil array to generate broadband electromagnetic waves. The transmitting coils at different positions can be controlled by using electric signals at different frequency ranges to generate broadband electromagnetic wave signals at different frequency ranges, so that the frequency range of channel detection is enlarged, and more detection information is obtained. The invention uses the emitting coil array to emit broadband electromagnetic wave as the detection signal, so that the detection signal passes through the PCB to be detected, and the geometric position of the emitting coil and the receiving signal can be accurately positioned at a certain detection position inside the PCB to be detected; by analyzing the array receiving signals and the array transmitting signals, the transfer characteristics of the PCB magnetic channels at the detection points can be inverted, and the frequency response characteristics of the PCB magnetic channels at different frequencies can be obtained. By way of example, for example: the frequency domain response characteristic of the received signal is obtained by frequency domain analysis, and is characterized by Y (x, Y, z; omega). Since the transmitted signal is a deterministic signal, it is characterized by X (X, y, z; ω). The channel characteristics of the detection PCB can be obtained according to the transmitting signals and the receiving signals, and the detection PCB is characterized by H (X, Y, z; omega), and the relation of the H (X, Y, z; omega) =Y (X, Y, z; omega)/X (X, Y, z; omega) is H (X, Y, z; omega). The (x, y, z) in the above formulas is a position coordinate, and the z-axis direction, namely the section characteristic of the PCB, can be well represented due to the adoption of an array for receiving and transmitting signals; ω is the angular frequency reflecting the response characteristics of the different frequencies. The time domain characteristics of the channel can be obtained by performing Fourier inverse transformation on H (x, y, z; omega), and can be represented by H (x, y, z; t), wherein t is time.
The invention also realizes a detection method for identifying the PCB based on the characteristics of the broadband magnetic induction channel, and the working flow chart is shown as 4, and comprises the following steps:
and 1, setting parameters of the device. The detection personnel carry out device parameter setting through the man-machine interaction module, and in the detection process, parameters needing to be set include:
(1) Setting transmission parameters, including waveform, center frequency, frequency range, strength and initial phase of the broadband signal, driving mode of the transmission coil, trend of current and power amplification factor selection;
(2) Setting receiving parameters, including offset value of frequency of a received signal, bandwidth of the signal, frequency and number of sampling and size of a receiving window;
(3) The setting of mechanical automation module parameters comprises the speed of belt transmission, time interval, action mode of a mechanical arm and time interval;
(4) The parameter setting of the infrared sensing counter comprises a correct counting initial value;
(5) Setting parameters of a communication module, including a communication connection mode and a network address;
(6) The parameter setting of data processing comprises a processing analysis method of signals in a channel inversion and fusion module, and algorithm types of feature comparison and classification judgment in a feature comparison module.
And 2, automatically detecting the state of the device. The detection content comprises detection of communication connection states between the front end and the data processing rear end, detection of connection states between the front end and each module of the data processing rear end, and supply states of power sources of the device. And only if the states detected by the device are normal, the next operation can be performed, otherwise, the alarm information is prompted on a display platform in the man-machine interaction module, and the detection flow is ended.
And step 3, importing a simulation data packet to establish a feature library. It is determined whether the model of PCB is first detected. If not, directly executing the step 4; if yes, execute this step. And importing the characteristic magnetic patterns, the characteristic parameters of the related magnetic patterns and the corresponding characteristic labels of the simulation system aiming at the detected PCB model into a database, and initially establishing a characteristic library of the PCB model.
And 4, emitting broadband detection electromagnetic waves. And generating a broadband signal according to the parameters set in the step 1. The broadband signal is converted into a broadband orthogonal electric signal through a digital-to-analog conversion unit and a power amplification unit in the preprocessing module. The broadband detection electromagnetic wave is generated by the electric signal acting on the transmitting coil array, and the control parameters of the transmitting coil array are also obtained by the setting of the step 1.
And 5, receiving an electromagnetic wave signal. The electromagnetic wave passing through the channel where the detected PCB is located is received by a receiving coil array, the parameters of the receiving coil array are set by the step 1, the received information is converted into digital information through an analog-to-digital conversion unit in a receiving buffer module and is stored in a buffer unit of the module, and the received electromagnetic wave information can be transmitted to the data processing back end through a communication module.
And 6, inverting and fusing the PCB magnetic channel characteristics. And the channel inversion and fusion module inverts the magnetic channel characteristics of the PCB at the detection point to obtain the magnetic induction intensity characteristics of the PCB at the detection point, and finally obtains the magnetic diagram of the PCB by splicing and fusing the magnetic induction intensity distribution characteristics at different positions.
And 7, comparing and classifying the magnetic channel characteristics. The feature comparison module directly calls the magnetic patterns in the PCB feature library of the same type in the database and corresponding features to compare, classify and judge, and the results can be divided into two types:
A. and determining a PCB detection result. The device can classify and judge the PCB through the characteristic comparison of the magnetic patterns, and the defect type and the position of the PCB are obtained. In this case, the detection result is directly recorded;
B. suspicious PCB detection results. The device can not accurately classify and judge the characteristics of the PCB magnetic diagram through the characteristic comparison of the magnetic diagram. In this case the PCB is marked and a warning is given. When the device is idle, manual detection or repeated detection can be selected. When in manual detection, not only the detection result of the PCB needs to be judged, but also the magnetic diagram of the detection PCB needs to be subjected to feature extraction, feature labels are built and put into a database, and sample features in the database are enriched. And then, a statistical mode recognition module, a neural network, a support vector machine and other machine learning algorithms are adopted, the classification judgment capacity is improved through training, and the judgment accuracy in an automatic detection mode is improved.
And 8, detecting the next PCB. Judging whether an instruction for ending the detection exists, and if so, ending the automatic detection. If not, the infrared counter counts up by 1, then the next PCB detection is carried out, and the step 4 is executed.
Claims (6)
1. The detection method based on the broadband magnetic induction channel characteristic identification PCB is characterized in that broadband electromagnetic waves are used as transmitting signals, the detected PCB is used as a channel, the array receiving signals and the array transmitting signals are utilized to invert the transfer function of the magnetic channel where the PCB is positioned at the detection point, the magnetic induction intensity of each point of the detected PCB is obtained, and then the spatial distribution diagram of the magnetic induction intensity characteristic of the detected PCB, namely a magnetic diagram, is obtained through splicing and fusion; the method comprises the following steps:
step 1, setting parameters of a device: setting device parameters by a detector through a man-machine interaction module;
step 2, automatic detection of device state: the detection content comprises a communication connection state between the detection front end and the data processing rear end, a connection state between each module of the detection front end and each module of the data processing rear end, a supply state of a device power supply, a next operation can be performed only under the condition that each state detected by the device is normal, otherwise, alarm information is prompted on a display platform in the man-machine interaction module, and the detection flow is ended;
step 3, importing a simulation data packet to establish a feature library: judging whether the PCB of the model is detected for the first time, if not, directly executing the step 4; if yes, executing the step, importing a data packet containing the characteristic magnetic pattern, the characteristic parameters of the related magnetic pattern and the characteristic label generated by the model PCB on the simulation system into a database, and initially establishing a characteristic library of the model PCB;
step 4, emitting broadband detection electromagnetic waves: generating a broadband signal according to the parameters set in the step 1; the broadband signal is converted into a broadband orthogonal electric signal through a digital-to-analog conversion unit and a power amplification unit in the preprocessing module, broadband detection electromagnetic waves are generated by the electric signal acting on the transmitting coil array, and control parameters of the transmitting coil array are also obtained by setting in the step 1;
step 5, receiving electromagnetic wave signals: the electromagnetic wave passing through the channel where the detected PCB is located is received by a receiving coil array, the parameters of the receiving coil array are set by the step 1, the received information is converted into digital information through an analog-to-digital conversion unit in a receiving buffer module and is stored in a buffer unit of the receiving buffer module, and the received electromagnetic wave information is transmitted to the data processing rear end through a communication module;
step 6, inversion and fusion of PCB magnetic channel characteristics: the channel inversion and fusion module inverts the magnetic channel characteristics of the PCB at the detection point to obtain the magnetic induction intensity characteristics of the PCB at the detection point, and finally obtains the magnetic diagram of the PCB by splicing and fusing the magnetic induction intensity distribution characteristics at different positions;
step 7, magnetic channel characteristic comparison and classification: the feature comparison module directly calls the magnetic patterns in the PCB feature library of the same type in the database and corresponding feature parameters to compare, classify and judge, and the results are divided into two types:
A. and (3) determining a PCB detection result: the PCB can be classified and judged through the characteristic comparison of the magnetic patterns, so that the defect type and the position of the PCB are obtained, and the detection result is directly recorded under the condition;
B. suspicious PCB detection results: the characteristics of the PCB magnetic pattern cannot be accurately classified and judged through the characteristic comparison of the magnetic pattern, in this case, the PCB is marked, warning prompt is provided, when the device is idle, manual detection or repeated detection is selected, when the manual detection is performed, the detection result of the PCB is required to be judged, the characteristic extraction is performed on the magnetic pattern of the detected PCB, a characteristic label is established and put into a database, the sample characteristics in the database are enriched, and then a machine learning algorithm is adopted, so that the classification judgment capability is improved through training, and the judgment accuracy in an automatic detection mode is improved;
step 8, detecting the next PCB: and judging whether an instruction for ending the detection exists, if so, ending the automatic detection, and if not, adding 1 to the count of the infrared counter, performing the next PCB detection, and executing the step 4.
2. The method for detecting the PCB based on the broadband magnetic induction channel feature identification according to claim 1, wherein the method is characterized in that a data packet comprising the simulated feature magnetic pattern, the related feature parameters and the classification labels of the PCB is imported to a database to establish a feature library during the first detection.
3. The detection method for identifying the PCB based on the broadband magnetic induction channel characteristics according to claim 1, wherein for suspicious results, a manual detection method is adopted to identify the defects of the PCB, and the automatic detection capability is improved through magnetic pattern characteristic extraction and sample training.
4. The detection method according to claim 1, wherein the machine learning method comprises one or more of statistical pattern recognition, neural network, and support vector machine.
5. The detection method according to claim 1, wherein the parameters to be set in step 1 include:
(1) Setting transmission parameters, including waveform, center frequency, frequency range, intensity and initial phase of broadband signals, driving mode of a transmission coil, trend of current and power amplification factor selection;
(2) Setting receiving parameters, including offset value of frequency of a received signal, bandwidth of the signal, frequency and number of sampling and size of a receiving window;
(3) The mechanical automation module parameters are set, including the speed and time interval of belt transmission, the action mode of the mechanical arm and the time interval;
(4) The parameter setting of the infrared sensing counter comprises a correct counting initial value;
(5) Setting parameters of a communication module, including a communication connection mode and a network address;
(6) Parameter setting of data processing: the method comprises a processing analysis method of signals in a channel inversion and fusion module, and an algorithm type of feature comparison and classification judgment in a feature comparison module.
6. The detection device for identifying the PCB based on the broadband magnetic induction channel characteristics for the detection method of claim 1, which is characterized by comprising a detection front end and a data processing rear end, wherein the detection front end is responsible for the transmission and the reception of broadband electromagnetic waves, and the data processing rear end is responsible for carrying out post-processing on a received signal to obtain a detection result; the detection front end comprises a slave controller, and a broadband signal generating module, a preprocessing module, a transmitting coil array, a receiving buffer module, an infrared induction counter, a mechanical automation module and a communication module which are respectively connected with the slave controller, wherein the broadband signal generating module, the preprocessing module and the transmitting coil array are sequentially connected, and the receiving coil array is connected with the receiving buffer module; the broadband signal generation module is mainly used for generating broadband detection signals, and the center frequency of the broadband detection signals is comprehensively considered and selected according to the detection depth and resolution requirements; the preprocessing module comprises a digital-to-analog conversion unit and a power amplification unit and is responsible for converting a digital signal into an electric signal; the transmitting coil array consists of a plurality of transmitting coils and transmits broadband electromagnetic wave signals under the action of electric signals; the receiving coil array consists of a plurality of coils and is used for receiving electromagnetic waves passing through the detected PCB channels at different positions; the receiving buffer module comprises an analog-to-digital conversion unit and a buffer unit, and is responsible for converting a received analog signal into a digital signal and storing the digital signal; the infrared induction counter is mainly used for counting the detected PCBs and is convenient for storage; the mechanical automation module comprises a belt transmission device and a mechanical arm and is mainly used for automatically placing and conveying the PCB; the communication module is mainly used for realizing communication with the data processing rear end; the slave controller is mainly used for carrying out parameter setting on each module in the detection front end and carrying out coordination control on the work among each module;
the data processing rear end comprises a database, a main controller, a man-machine interaction module, a channel inversion and fusion module, a characteristic comparison module, a data storage module and a communication module, wherein the man-machine interaction module, the channel inversion and fusion module, the characteristic comparison module, the data storage module and the communication module are respectively connected with the main controller; the human-computer interaction module is mainly used for configuring parameters of the whole device, importing feature data detected for the first time and carrying out information interaction of manual detection; the database is used for storing a characteristic library and comprises a characteristic magnetic diagram of the PCB to be detected, related characteristic parameters and characteristic labels; the channel inversion and fusion module inverts the channel of the PCB at the detected point by utilizing the transmitting signal and the receiving signal to obtain the magnetic induction intensity of the PCB at the detected point, and then the magnetic induction intensity characteristic distribution map of the detected PCB is obtained by splicing and fusing the magnetic induction intensities of all points; the feature comparison module is used for extracting features of the detected PCB magnetic patterns, comparing the feature extraction with features in a PCB feature library of the same type in the database to obtain a feature classification result, and judging the cause and the position of the defect; the data storage module is used for storing the detection result and the corresponding detection count number of the detected PCB; the communication module is mainly used for realizing communication with the detection front end; the main controller is mainly used for setting parameters of the system, and is used for coordinately controlling the function realization of each module and the mutual calling among the modules.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710336742.5A CN107290653B (en) | 2017-05-13 | 2017-05-13 | Detection device and method for identifying PCB based on broadband magnetic induction channel characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710336742.5A CN107290653B (en) | 2017-05-13 | 2017-05-13 | Detection device and method for identifying PCB based on broadband magnetic induction channel characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107290653A CN107290653A (en) | 2017-10-24 |
| CN107290653B true CN107290653B (en) | 2023-04-21 |
Family
ID=60094523
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710336742.5A Active CN107290653B (en) | 2017-05-13 | 2017-05-13 | Detection device and method for identifying PCB based on broadband magnetic induction channel characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107290653B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108229672B (en) * | 2018-02-13 | 2022-03-29 | 北京旷视科技有限公司 | Processing method, device and equipment of deep neural network model |
| CN108878095A (en) * | 2018-09-01 | 2018-11-23 | 浙江行雨网络科技有限公司 | A kind of dynamic degaussing gear of unattended supermarket self-help settlement commodity |
| CN109919925A (en) * | 2019-03-04 | 2019-06-21 | 联觉(深圳)科技有限公司 | Printed circuit board intelligent detecting method, system, electronic device and storage medium |
| CN110794353A (en) * | 2019-11-05 | 2020-02-14 | 海南电网有限责任公司电力科学研究院 | Method for simulating broadband large current and broadband large current frequency sweep test system |
| CN112578264A (en) * | 2020-11-25 | 2021-03-30 | 海鹰企业集团有限责任公司 | Positioning device for various short circuit points of circuit board card |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104655714A (en) * | 2014-12-05 | 2015-05-27 | 广州丰谱信息技术有限公司 | Detecting and imaging method and detecting and imaging device based on broadband magnetic wave reflection path parameter identification |
| CN104655656A (en) * | 2014-12-05 | 2015-05-27 | 广州丰谱信息技术有限公司 | Detection imaging method and detection imaging device based on broadband magnetic wave transmission model parameter identification |
| CN105380645A (en) * | 2015-10-26 | 2016-03-09 | 华南理工大学 | Detection method and device for magnetopneumogram |
| WO2017016080A1 (en) * | 2015-07-24 | 2017-02-02 | 广州彩磁信息技术有限公司 | Security inspection system and method based on broadband detection and visual display using conjugated electromagnetic transmitting and receiving arrays |
-
2017
- 2017-05-13 CN CN201710336742.5A patent/CN107290653B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104655714A (en) * | 2014-12-05 | 2015-05-27 | 广州丰谱信息技术有限公司 | Detecting and imaging method and detecting and imaging device based on broadband magnetic wave reflection path parameter identification |
| CN104655656A (en) * | 2014-12-05 | 2015-05-27 | 广州丰谱信息技术有限公司 | Detection imaging method and detection imaging device based on broadband magnetic wave transmission model parameter identification |
| WO2017016080A1 (en) * | 2015-07-24 | 2017-02-02 | 广州彩磁信息技术有限公司 | Security inspection system and method based on broadband detection and visual display using conjugated electromagnetic transmitting and receiving arrays |
| CN105380645A (en) * | 2015-10-26 | 2016-03-09 | 华南理工大学 | Detection method and device for magnetopneumogram |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107290653A (en) | 2017-10-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107290653B (en) | Detection device and method for identifying PCB based on broadband magnetic induction channel characteristics | |
| Bernieri et al. | Crack shape reconstruction in eddy current testing using machine learning systems for regression | |
| US20200300923A1 (en) | Automatic test system of wireless charging system | |
| CN107678960B (en) | Intelligent ammeter software black box testing system and method | |
| CN105022020B (en) | Electric energy meter radio frequency electromagnetic field, radio immunity experiment test system and method | |
| CN102187233B (en) | Measuring system for determining scatter parameters | |
| CN103071909A (en) | Real-time detection and weld quality prediction in vibration welding process | |
| CN106772287B (en) | A kind of Radar Automatic Test System of generic Extensible | |
| CN105675115B (en) | A kind of online vibration measuring system of laser-Doppler and method | |
| CN103760224A (en) | Spacecraft nondestructive detection sensor and nondestructive detection system | |
| CN103995202B (en) | A kind of automatic signal method of testing, Apparatus and system | |
| CN104655656B (en) | Detection imaging method and device based on wideband magnetic wave transmission model parameter identification | |
| CN206649137U (en) | A kind of anti-interference monitoring system of car bulb | |
| CN203433066U (en) | Electrical property automatic detection device of piezoelectric ceramic buzzer | |
| CN105241951A (en) | Non-magnetic conductor material electromagnetic eddy current detection apparatus | |
| CN104390982A (en) | Test method for SMT first article inspection | |
| CN208984725U (en) | A kind of system detecting capacitive touch screen functional sheet | |
| CN109239469B (en) | Device and method for detecting magnetic shielding effect of magnetic shielding material under low magnetic field | |
| CN103983887B (en) | Automatic locating method and device for open circuit of circuit | |
| CN105652188A (en) | Contactless card radio frequency performance automatic test system and test method thereof | |
| CN110031511A (en) | Defect detecting device, defect detecting system and defect inspection method | |
| CN109283500B (en) | Online testing system and method for mass production radar | |
| CN103760226A (en) | Non-destructive detection method of spacecraft | |
| CN107121624A (en) | Partial discharge detection multi-drop arrangement and localization method based on moment method | |
| CN208156123U (en) | A kind of antenna Board Test System |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |